Algorithms (Secretly) Run the World

Megalith

24-bit/48kHz
Joined
Aug 20, 2006
Messages
13,000
Are you in control of your life? Maybe not, or, at least, not as much as you think. See, complex mathematical formulas are basically intertwined with anything and everything these days, and no one is batting an eye at how many decisions are essentially being made for you. Of course, the problem here is that algorithms are not perfect—more specifically, they originated with a human touch, meaning that they are arguably tainted with people problems. Er, so what is the solution here, considering that “bias, accountability and ethical decisions...exist whether you are using algorithms or not”? Thanks to Kyle for this one.

When you browse online for a new pair of shoes, pick a movie to stream on Netflix or apply for a car loan, an algorithm likely has its word to say on the outcome. The complex mathematical formulas are playing a growing role in all walks of life: from detecting skin cancers to suggesting new Facebook friends, deciding who gets a job, how police resources are deployed, who gets insurance at what cost, or who is on a "no fly" list. …fears are rising over the lack of transparency algorithms can entail, with pressure growing to apply standards of ethics or "accountability." Data scientist Cathy O'Neil cautions about "blindly trusting" formulas to determine a fair outcome. "Algorithms are not inherently fair, because the person who builds the model defines success," she said.
 
Click-bait hype.

Algorithms can be complex, but they don't have to be complex. Furthermore algorithms only do what they have been written to do, the "choices" are not made by the algorithm but by the writer, or more accurately, the entity that the algorithm is "working for". The idea that the algorithm is making decisions for you is stupid, it's just doing what it was written to do, the choices are based on the policies and rules set in place by the business that is using it.

That being said, a good algorithm is going to work properly and a bad one ...... well.

O'Neil argues that while some algorithms may be helpful, others can be nefarious. In her 2016 book, "Weapons of Math Destruction," she cites some troubling examples in the United States:

- Public schools in Washington DC in 2010 fired more than 200 teachers -- including several well-respected instructors -- based on scores in an algorithmic formula which evaluated performance.

- A man diagnosed with bipolar disorder was rejected for employment at seven major retailers after a third-party "personality" test deemed him a high risk based on its algorithmic classification.

These examples from the article are not examples of "bad" algorithms, they are examples of "possibly" bad decisions based upon algorithmic data.
 
Click-bait hype.

Algorithms can be complex, but they don't have to be complex. Furthermore algorithms only do what they have been written to do, the "choices" are not made by the algorithm but by the writer, or more accurately, the entity that the algorithm is "working for". The idea that the algorithm is making decisions for you is stupid, it's just doing what it was written to do, the choices are based on the policies and rules set in place by the business that is using it.

That being said, a good algorithm is going to work properly and a bad one ...... well.



These examples from the article are not examples of "bad" algorithms, they are examples of "possibly" bad decisions based upon algorithmic data.
Well the initial blurb did say fears are rising over the lack of transparency algorithms can entail, with pressure growing to apply standards of ethics or "accountability."

If the algorithms are not published, then they could be doing anything. For instance, there could be a line "if subject = bipolar then hire=false" which would be worrying for anyone listing themselves as bipolar.
 
And if bias by the developers of the algorithms isn't bad enough, there are the implementation (coding) bugs that we have to be concerned about too. There's a lot of software being slapped together out there by people with little talent for the job, exacerbated by poor or non-existent testing protocols.
 
Well the initial blurb did say fears are rising over the lack of transparency algorithms can entail, with pressure growing to apply standards of ethics or "accountability."

If the algorithms are not published, then they could be doing anything. For instance, there could be a line "if subject = bipolar then hire=false" which would be worrying for anyone listing themselves as bipolar.

OK, but the algorithm is a tool used to help with decision making correct?

But it doesn't make the decision, but someone does have to tell the tool what decisions to help with or how to help.

As for transparency, if you didn't have a "transparent" algorithm, then how would you prove discrimination? And if you do have an algorithm, why can't the same still work?
 
I think that's the point.


I don't understand what this means.

What I am saying is the discrimination has been an issue long before computers. And discrimination law suites were being settled without them as well. So even if a company uses an algorithm tool to assist with hiring, and intentional or unintentionally, this tool is effectively causing discrimination. It can still be identified and proven in the same old tried and true manner. The fact that an algorithm is involved is immaterial in identifying discrimination, although it could help point out whether the discrimination was intentional or not.

I just don't think you need oversight and inspection of such things. The tool isn't the problem, it's the people behind the tool either through malfeasance or incompetence, it's still down to people. But the results remain observable or indicative of a problem if a problem exists to begin with.
 
What I am saying is the discrimination has been an issue long before computers. And discrimination law suites were being settled without them as well. So even if a company uses an algorithm tool to assist with hiring, and intentional or unintentionally, this tool is effectively causing discrimination. It can still be identified and proven in the same old tried and true manner. The fact that an algorithm is involved is immaterial in identifying discrimination, although it could help point out whether the discrimination was intentional or not.

I just don't think you need oversight and inspection of such things. The tool isn't the problem, it's the people behind the tool either through malfeasance or incompetence, it's still down to people. But the results remain observable or indicative of a problem if a problem exists to begin with.
I don't really think this only relates to lawsuits and discrimination, but that too many people think of systems as magical boxes that always have the best answers when reality can be quite different. A problem with releasing algorithms would be that people will know exactly how to game the systems and obtain optimal results.
 
Back
Top